Estimation of frequency difference at which stream segregation precedes temporal integration as reflected by omission mismatch negativity.

Both stream segregation and temporal integration are considered important for auditory scene analysis in the brain. Several previous studies have indicated that stream segregation may precede temporal integration when both processes are required. In the present study, we utilized mismatch negativity (MMN)-which reflects automatic change detection-to systematically estimate the threshold of the frequency difference at which stream segregation occurs prior to temporal integration when these functions occur together during a state of inattention. Electroencephalography (EEG) data were recorded from 22 healthy Japanese men presented with six blocks of alternating high pure tones (high tones) and low pure tones (low tones). Only high tones were omitted with 5 % probability in all blocks. Our results indicated that stream segregation should cancel temporal integration of close sounds, as indicated by omission-MMN elicitation, when the frequency difference is 1000 Hz or larger.